GitHub链接:DeepLearnToolbox DeepLearnToolbox A Matlab toolbox for Deep Learning. Deep Learning is a new subfield of machine learning that focuses on learning deep hierarchical models of data. It is inspired by the human brain's apparent deep (layered,
前言: CNN作为DL中最成功的模型之一,有必要对其更进一步研究它.虽然在前面的博文Stacked CNN简单介绍中有大概介绍过CNN的使用,不过那是有个前提的:CNN中的参数必须已提前学习好.而本文的主要目的是介绍CNN参数在使用bp算法时该怎么训练,毕竟CNN中有卷积层和下采样层,虽然和MLP的bp算法本质上相同,但形式上还是有些区别的,很显然在完成CNN反向传播前了解bp算法是必须的.本文的实验部分是参考斯坦福UFLDL新教程UFLDL:Exercise: Convolutional Ne
最近在学深度学习相关的东西,在网上搜集到了一些不错的资料,现在汇总一下: Free Online Books by Yoshua Bengio, Ian Goodfellow and Aaron Courville Neural Networks and Deep Learning42 by Michael Nielsen Deep Learning27 by Microsoft Research Deep Learning Tutorial23 by LISA lab, University
Awesome Deep Learning Table of Contents Free Online Books Courses Videos and Lectures Papers Tutorials Researchers WebSites Datasets Frameworks Miscellaneous Contributing Free Online Books Deep Learning by Yoshua Bengio, Ian Goodfellow and Aaron Cou
前言: 当采用无监督的方法分层预训练深度网络的权值时,为了学习到较鲁棒的特征,可以在网络的可视层(即数据的输入层)引入随机噪声,这种方法称为Denoise Autoencoder(简称dAE),由Bengio在08年提出,见其文章Extracting and composing robust features with denoising autoencoders.使用dAE时,可以用被破坏的输入数据重构出原始的数据(指没被破坏的数据),所以它训练出来的特征会更鲁棒.本篇博文主要是根据Benig